Akademik Veri Yönetim Sistemi
STRUCTURAL CONCRETE, cilt.27, sa.2, ss.1-25, 2026 (SCI-Expanded, Scopus)
Lightweight strain-hardening ultra-high-performance concrete composite(SH-UHPC) is an outstanding alternative for engineering applications andinfrastructure thanks to its outstanding strength, toughness, ductility, and lowdensity. The integration of artificial intelligence (AI)-based modeling strategiesinto engineering problems can substantially accelerate material design pro-cesses while reducing experimental costs and time. Within this scope, the mainmotivation of this study is to predict the compressive strength (CS) of light-weight SH-UHPC via a grey wolf optimization (GWO)-integrated machinelearning (ML)-based modeling that offers high accuracy and reliability, therebyreducing experimental cost and time requirements while supporting environ-mental and economic sustainability. The overall results demonstrate that alldeveloped GWO-ML models achieved impressive performance levels in pre-dicting the CS of lightweight SH-UHPC. In particular, the GWO-Extra TreesRegressor (GWO-ETR) model demonstrated superior performance comparedto other GWO-ML models in terms of performance metrics (RMSE = 6.99,MAPE = 3.25%, and R2 = 0.929), scatter plots (all points remained within a10% margin of error), and uncertainty analysis (U95 = 10.0) during the testingphase. In addition, SHapley Additive exPlanations-based feature importanceanalysis, as well as individual conditional expectation analysis and partialdependence plots, provided valuable insights and design suggestions for engi-neers and practitioners. Finally, an interactive graphical user interface hasbeen developed to enable the application of similar data-driven analyses on alarger scale and to obtain rapid predictions; however, it is suggested that thedatabase be continuously updated to improve model performance and extendits generalization capacity in the future.